America’s ageing economic boom can still produce pleasant surprises. Companies added an astonishing 312,000 new jobs in December, the Bureau of Labour Statistics reported on January 4th, and raised pay at the fastest clip in years. For the third of working-age Americans without any college education, such spells of rapid income growth have been exceedingly rare, not only since the financial crisis but in the past half-century. But however long this expansion lasts, their economic prospects still look grim.
The misfortunes of the left-behind were a recurring topic at this year’s meeting, in Atlanta, of the American Economic Association, one of the biggest annual convocations of economists. David Autor of the Massachusetts Institute of Technology offered the most pointed characterisation, drawing on forthcoming research co-written with Juliette Fournier, also of mit. The earnings of workers without a college education have scarcely risen in 50 years, after adjusting for inflation; for men they have fallen. This stagnation coincided with tectonic changes in American employment. The share of jobs that require either a lot of training, or very little, has grown since 1970. Much of the production and office work that requires moderate training, which once employed vast numbers of workers without college degrees, has disappeared, either shipped abroad or offloaded on robots and computers. The resulting hardship has been implicated in a rise in mortality in parts of America and the turn toward angry nationalism that helped put Donald Trump in the White House.
Working out what to do about those left behind by economic progress is becoming an obsession of policy wonks. Mr Autor and Ms Fournier provide important new context. In the 1950s, they show, there was almost no relationship between how densely populated a place was and the share of its residents with college degrees. That has changed utterly: the share of the working-age population with a college degree is now 20 percentage points higher in urban places than it is in rural ones. In 1970 that gap was just five percentage points. Several decades ago mid-skilled work was clustered in big cities, while low-skilled work was most prevalent in the countryside. No longer; those mid-skilled jobs that remain are more likely to be found in rural areas than in urban ones.
As the geographical pattern of work has shifted, so has that of wages. Economists have long acknowledged the existence of an urban wage premium: workers in more densely populated places earn more, in part because of the productivity benefits of crowding together that nurture urban growth in the first place. This pay premium used to hold across the range of skills. In 1970 workers without any college education could expect to get a boost to their earnings when they moved to a big city, just as better-educated workers did (see chart). Since then the urban wage advantage for well-educated workers has become more pronounced, even as that for less-educated workers has all but disappeared….